Fig. 4
From: Fast estimation of plant growth dynamics using deep neural networks
Plotted are the results of a linear model (estimate and 95% confidence bands) with the logarithm of the distance between user labelled points and predicted points as a response variable and the interaction of plant part and the number of labelled training frames the models as explanatory variables. For this analysis plant parts were categorized to make them comparable across the different video analysis, where B (base), M (mid-section), T (tip) and TB (tip branching) represent the label categories detailed in Table 2. Note that due to the different image sizes across videos, distances cannot be directly compared across analyses in this plot